*** rlandy is now known as rlandy|bbl | 00:04 | |
clarkb | corvus: looks like NS zuul-ci.org is an empty response currently | 00:53 |
---|---|---|
*** smyers has quit IRC | 01:07 | |
*** tflink has quit IRC | 01:08 | |
*** smyers has joined #zuul | 01:08 | |
*** tflink has joined #zuul | 01:10 | |
fungi | the glue records seem to be missing in the org tld | 01:14 |
fungi | they have the ns records but not the a/aaaa glue to go with them | 01:15 |
fungi | our nameservers are serving both the ns and a/aaaa records for them just fine as far as i can tell | 01:15 |
clarkb | fungi: if I dig NS I don't get an NS record at all | 01:16 |
*** tflink has quit IRC | 01:17 | |
fungi | dig +trace zuul-ci.org @a.root-servers.net. | 01:17 |
fungi | you'll see 'em | 01:17 |
clarkb | huh dig must only show you NS records if it can lookup a/aaaa by default? | 01:18 |
clarkb | but it doesn't show you the a/aaaa values | 01:18 |
fungi | right | 01:18 |
fungi | the org nameservers are supposed to be including complimentary a/aaaa records for them, called "glue" records | 01:18 |
fungi | but seem to be missing them | 01:19 |
clarkb | right, I just expected dig to return an NS record if the .org server had them | 01:20 |
clarkb | and it doesn't | 01:20 |
fungi | a temporary workaround would be for corvus to change the nameservers on the domain to ns1/ns2.openstack.org which can resolve without zuul-ci.org depending on its own resolution | 01:21 |
fungi | (and resolves to the same servers as ns1/ns2.zuul-ci.org is supposed to) | 01:21 |
fungi | i have a suspicion what bramwelt saw with zuul-ci.org resolution back at 17:58 utc was related | 01:26 |
dmsimard | Unimportant question: should the project url in github for zuul be zuul-ci.org instead of openstack.org ? | 01:27 |
*** tflink has joined #zuul | 01:27 | |
fungi | what url in github? | 01:29 |
dmsimard | https://github.com/openstack-infra/zuul, ctrl+f openstack.org | 01:36 |
fungi | ahh, you mean in the github repository description field | 01:40 |
dmsimard | oh, that's what it's called | 01:40 |
dmsimard | ¯\_(ツ)_/¯ | 01:40 |
fungi | ahh, in the updated webui you click the "edit" button and it shows those as separate "description" and "url" fields | 01:41 |
fungi | it just displays them jammed together | 01:42 |
fungi | so what you said was an apropos name | 01:42 |
mordred | I'd say 'yes' - the project url should be zuul-ci.org | 01:45 |
*** rlandy|bbl is now known as rlandy | 01:54 | |
clarkb | just dont chamge it until dns works :) | 01:59 |
fungi | indeed, but also it's in the openstack org on gh anyway | 02:05 |
fungi | as part of the winterscale effort, i expect we'll be able to consider all manner of options (including possibly not mirroring it to gh any longer?) | 02:06 |
mordred | fungi: yah - I think there are many things we can/should consider related to mirroring and winterscale and whatnot | 02:11 |
openstackgerrit | Merged openstack-infra/zuul-jobs master: Add guidelines on variable naming https://review.openstack.org/595920 | 02:12 |
dmsimard | oops, got another report that zuul-ci.org was unreachable.. could it be the DNS issue ? | 02:40 |
dmsimard | works for me | 02:41 |
dmsimard | ¯\_(ツ)_/¯ | 02:41 |
clarkb | yes, this is the dns issue | 02:41 |
*** rlandy has quit IRC | 03:13 | |
dmsimard | I suck at graphs and I'm presenting something tomorrow. Would you say this is an accurate portrayal of a high level view of the Zuul workflow ? https://i.imgur.com/L7Q36vJ.png | 04:00 |
dmsimard | URL if you want to toy with it: http://paste.openstack.org/show/729573/ | 04:01 |
dmsimard | It's probably not all 100% true due to lack of depth/accuracy, the emphasis is more on the "what" than the "how". Feedback still appreciated :) | 04:03 |
dmsimard | tweaked a few things around the nodepool node creation, newer version available here: http://paste.openstack.org/show/729574/ | 04:10 |
dmsimard | going to sleep :p | 04:11 |
openstackgerrit | Simon Westphahl proposed openstack-infra/nodepool master: Implement liveness probe for static driver https://review.openstack.org/600083 | 05:18 |
openstackgerrit | Simon Westphahl proposed openstack-infra/nodepool master: Cleanup of leaked resource for static driver https://review.openstack.org/600084 | 05:18 |
*** hashar has joined #zuul | 05:35 | |
openstackgerrit | Merged openstack-infra/zuul-jobs master: Remove trigger-readthedocs job https://review.openstack.org/583836 | 05:38 |
openstackgerrit | Merged openstack-infra/zuul master: Remove restriction on add_host https://review.openstack.org/600205 | 05:59 |
*** pcaruana has joined #zuul | 06:25 | |
*** goern has quit IRC | 06:38 | |
*** sshnaidm|off is now known as sshnaidm | 07:06 | |
*** ssbarnea|bkp2 has quit IRC | 07:27 | |
*** goern has joined #zuul | 07:34 | |
*** pcaruana has quit IRC | 07:36 | |
*** ssbarnea|bkp has joined #zuul | 07:40 | |
*** pcaruana has joined #zuul | 07:48 | |
*** jpena|off is now known as jpena | 07:49 | |
*** sshnaidm is now known as sshnaidm|afk | 08:31 | |
*** electrofelix has joined #zuul | 08:48 | |
*** hwoarang_ has joined #zuul | 09:11 | |
*** hwoarang has quit IRC | 09:15 | |
*** sshnaidm|afk is now known as sshnaidm | 09:16 | |
*** hwoarang_ is now known as hwoarang | 09:23 | |
*** andreaf has joined #zuul | 09:47 | |
*** panda has quit IRC | 09:48 | |
*** hashar is now known as hasharAway | 09:53 | |
ianw | urgh, getting dns failures to zuul-ci.org | 10:07 |
*** jpena is now known as jpena|lunch | 11:02 | |
*** swest has quit IRC | 11:23 | |
*** swest has joined #zuul | 11:23 | |
*** swest has quit IRC | 11:28 | |
*** swest has joined #zuul | 11:42 | |
*** hasharAway is now known as hashar | 11:49 | |
*** rfolco has joined #zuul | 11:53 | |
*** jpena|lunch is now known as jpena | 11:59 | |
*** rlandy has joined #zuul | 12:38 | |
openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: WIP: Match tag items against containing branches https://review.openstack.org/578557 | 13:08 |
tobiash | corvus: I hope you don't mind if I continue on ^ | 13:09 |
fungi | ianw: yeah, i think the issue is progressively hitting people as ns records age out of recursive dns caches | 13:32 |
fungi | once corvus is around he may be able to either figure out why gandi is no longer injecting those glue records into the org tld, or temporarily switch the nameservers of record to another name we have resolving for them | 13:34 |
*** samccann has joined #zuul | 13:35 | |
openstackgerrit | Fabien Boucher proposed openstack-infra/zuul-jobs master: ensure-sphinx: do not attempt a gettext install if exists https://review.openstack.org/599028 | 13:42 |
*** sshnaidm is now known as sshnaidm|mtg | 13:49 | |
*** hashar is now known as hasharAway | 13:59 | |
pabelanger | so, I'm not sure how best to solve the issue where a zuul-executor needs an ansible dependency to run a task, for example docker-py: http://logs.openstack.org/70/592770/12/check/ansible-role-nodepool-docker-ubuntu-bionic/2f7356d/job-output.txt.gz#_2018-09-06_01_53_48_429332 | 14:23 |
pabelanger | right now, we cannot use any of the docker modules from zuul-executor in openstack-infra, and unsure if pip installing for all jobs is worth it, or some sort of dynamic dependency list at run time | 14:24 |
pabelanger | I see this kinda falling into the same place where job could change which version of ansible they are using | 14:25 |
dmsimard | pabelanger: yeah, it's in the same bucket because there are a lot of dependencies that aren't installed by default | 14:25 |
dmsimard | docker-py is one but there's openstacksdk for openstack modules, boto for aws modules, etc | 14:26 |
pabelanger | yah | 14:26 |
openstackgerrit | Markus Hosch proposed openstack-infra/zuul master: Add --check-config option to zuul scheduler https://review.openstack.org/542160 | 14:51 |
corvus | fungi: i'm a bit confused. this appears to return the correct data: dig +dnssec zuul-ci.org. a @b2.org.afilias-nst.org. | 15:11 |
corvus | fungi: i repeated that for all of the .org servers | 15:12 |
corvus | fungi: (i agree that +trace fails. though +trace apparently never shows the additional data anyway?) | 15:13 |
fungi | corvus: yeah, that is really strange :/ | 15:13 |
fungi | and +additional seems to show it | 15:18 |
fungi | which makes this even more confounding | 15:18 |
fungi | dig +trace +additional zuul-ci.org a @a.root-servers.net. | 15:18 |
corvus | yeah -- "couldn't get address for ..." right below the address for it | 15:19 |
fungi | probably unrelated, reverse dns is missing on the ns2.o.o ipv6 addy | 15:20 |
*** sshnaidm|mtg is now known as sshnaidm | 15:20 | |
clarkb | corvus: supposedly dig uses local resolver to resolve that step rather than what was just returned? | 15:26 |
corvus | could it be that our dnssec keys are expired? | 15:30 |
*** ssbarnea|bkp has quit IRC | 15:32 | |
corvus | fungi, clarkb: the expiration of the rrsig field for ns1 is 20180904222910 | 15:37 |
fungi | huh... | 15:38 |
fungi | indeed it is | 15:39 |
fungi | generated 20180805222718 ? | 15:39 |
fungi | /var/lib/bind/zones/zuul-ci.org/zone.db.signed on adns1 was last modified 5 days ago | 15:40 |
fungi | did it not propagate? | 15:41 |
corvus | Sep 6 15:27:24 ns2 nsd[22004]: zuul-ci.org: Could not tcp connect to 23.253.63.149: No route to host | 15:42 |
corvus | (that ip is ads1) | 15:42 |
fungi | confirmed | 15:42 |
*** pcaruana has quit IRC | 15:42 | |
fungi | looks like we broke iptables on adns1? | 15:43 |
corvus | did we bust iptables on adns1 with the ansible stuff? | 15:43 |
fungi | i believe so | 15:43 |
corvus | shall we move this to #openstack-infra now? | 15:44 |
fungi | yes | 15:44 |
*** pcaruana has joined #zuul | 16:01 | |
*** hasharAway is now known as hashar | 16:08 | |
*** sshnaidm is now known as sshnaidm|off | 16:10 | |
*** ssbarnea|bkp has joined #zuul | 16:14 | |
*** jpena is now known as jpena|off | 16:36 | |
*** ChanServ sets mode: -rf | 17:12 | |
*** bhavikdbavishi has joined #zuul | 17:21 | |
*** pcaruana has quit IRC | 17:27 | |
*** electrofelix has quit IRC | 17:59 | |
*** pcaruana has joined #zuul | 18:02 | |
*** openstackgerrit has quit IRC | 18:05 | |
fungi | looks like zuul-ci.org dns resolution should be fixed now. anyone experiencing resolution issues, please let us know in here or in #openstack-infra | 18:08 |
*** _ari__ has joined #zuul | 18:46 | |
*** bhavikdbavishi has quit IRC | 18:51 | |
*** AJaeger_ has joined #zuul | 19:00 | |
*** AJaeger has quit IRC | 19:02 | |
*** AJaeger_ is now known as AJaeger | 19:03 | |
dmsimard | fungi, corvus, clarkb: any thoughts on https://i.imgur.com/z0fKZM7.png ? Was part of a short demo I ran this morning. If it's useful and accurate enough I can provide the editable source (draw.io XML export) | 19:07 |
AJaeger | zuul experts, could you comment on https://review.openstack.org/#/c/591528/7/.zuul.yaml , please? I think we don't need the first job and can just use the parent - correct? | 19:10 |
fungi | dmsimard: do you think that would be useful on the website? in the docs? on marketing materials? what was the target audience? | 19:11 |
clarkb | reading it I expect that the target audience is someone that is trying to understand how zuul works (so someone deploying zuul or evaluating it for deplyoment) | 19:11 |
clarkb | it seems accurate too | 19:11 |
fungi | dmsimard: what about including the git driver? i expect that would basically just go straight from human to zuul scheduler | 19:12 |
fungi | with the line labelled something like "merges commit" | 19:13 |
dmsimard | fungi: clarkb beat me to the punch, the objective was to explain the zuul process/workflow at a high level to technical people familiar with Jenkins | 19:13 |
fungi | though i do think the chart is useful even without that | 19:13 |
fungi | but a lot of people familiar with jenkins likely use it without direct involvement with code review, so maybe the git driver does make sense to them too, i dunno | 19:14 |
dmsimard | fungi: I admittedly had a bit of a struggle trying to convey a good level of detail in the graph without including everything which I felt would make it harder to understand | 19:16 |
fungi | yep, it looks remarkably comprehensive to me | 19:17 |
fungi | dmsimard: one question on the not pictured bullets... do we actually have elk log submission/classification driven from the zuul-jobs stdlib? | 19:18 |
dmsimard | ah, true, I guess that's openstack-infra and software factory specific | 19:19 |
corvus | dmsimard: very nice | 19:27 |
AJaeger | corvus, could you comment on https://review.openstack.org/#/c/591528/7/.zuul.yaml , please? I think we don't need the first job and can just use the parent - correct? | 19:28 |
corvus | AJaeger: was just looking at that | 19:28 |
corvus | AJaeger: it would be nice of those jobs had 'description's | 19:29 |
AJaeger | yeah ;( | 19:30 |
AJaeger | thanks, corvus - if you have questions ask - I dug a bit into those already | 19:31 |
corvus | AJaeger: let's move to infra | 19:31 |
AJaeger | ok | 19:31 |
*** pcaruana has quit IRC | 19:35 | |
*** johanssone has joined #zuul | 19:59 | |
*** elyezer__ has joined #zuul | 20:27 | |
*** elyezer_ has quit IRC | 20:30 | |
*** openstackgerrit has joined #zuul | 20:41 | |
openstackgerrit | Merged openstack-infra/zuul-jobs master: Log output and exception when traceroute fails https://review.openstack.org/576351 | 20:41 |
*** rfolco has quit IRC | 21:00 | |
*** rfolco has joined #zuul | 21:00 | |
*** samccann has quit IRC | 21:08 | |
*** sshnaidm|off has quit IRC | 21:36 | |
dmsimard | corvus: I guess my question was if you had a use for it :p | 21:44 |
corvus | dmsimard: i feel like there should be a place in the docs for it | 21:47 |
dmsimard | It's probably tutorial/introduction/zuul from scratch content ? | 21:48 |
corvus | dmsimard: i think introductory; maybe one of the first 2 pages of the user guide? or somewhere near there? | 21:48 |
corvus | dmsimard: probably not zfs where i think we want to be focused just on "how do you get this thing going?" | 21:49 |
corvus | dmsimard: or maybe it should be in the admin guide instead of user? | 21:49 |
corvus | dmsimard: there's some definite user/admin crossover going on there :) | 21:49 |
dmsimard | Well, there's nothing preventing us from having different versions of it for different purposes | 22:24 |
*** hashar has quit IRC | 22:29 | |
*** jimi|ansible has quit IRC | 22:30 | |
*** jimi|ansible has joined #zuul | 22:40 | |
*** jimi|ansible has joined #zuul | 22:40 | |
fungi | or from including the same diagram in two different pages. it's not like we need to ship multiple copies to make that happen | 22:45 |
*** jimi|ansible has quit IRC | 23:42 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!